Sixteen million people in the United States live with the chronic disease diabetes; approximately 5 -10% are children. The seminal Diabetes Control & Complications Trial (DCCT) concluded that frequent glucose monitoring is necessary to reduce the complications of this disease. However, all glucose monitors available require invasive techniques with the most widely used method of self-monitoring, obtaining blood from a finger stick, causing pain and discomfort which results in poor compliance. The development of a novel, hand-held, non-invasive glucose monitor will provide diabetics with the means for testing their glucose level more frequently, improving their quality of life, and reducing the costs and complications of this chronic disease. This application proposes the development, calibration and clinical validation of an advanced prototype based on a proprietary method of mid-infrared Thermal Emission Spectroscopy (TES). The method and instrument are based on the discovery that natural mid-infrared emissions, from the tympanic membrane, consist of spectral information of blood analytes. In recent years, infrared (IR) spectroscopy has emerged as a potential analytical method of choice for non-invasive glucose monitoring. The specific aims of this Phase I proposal are to develop the optical system and to integrate sensors into an ergonomic form for diabetic self-monitoring. The subsequent clinical studies are designed to demonstrate that a universal calibration for real time accurate measurements is possible for predictive results. The device will be calibrated and validated across a range of glucose concentrations from 40 to 400 mg/dL, using a total of 52 diabetic and normal subjects in hyperglycemic and hypoglycemic protocols. A trained nurse will perform the measurements. Subsequently, 19 diabetic subjects will test themselves using the hyperglycemic protocol in a full predictive study. The hypotheses will test that the intra-individual correlation coefficient of the non-invasive predicted serum glucose compared with the clinical laboratory serum glucose measurement will be greater than 0.85, and less than 1 percent of the measurements will fall outside the clinically acceptable range on the Clarke Error Grid. Successful conclusion of Phase I milestones will lead to Phase II clinical studies and modifications for younger pediatric subjects.